Table of Content

15 December 2022, Volume 26 Issue 4

Previous Issue    Next Issue

Research on Black-Litterman portfolio model based on financial text sentiment mining—Evidence from the posting text of eastmoney stock forum and the A share market

XU Weijun, HUANG Jinglong, FU Zhineng, ZHANG Weiguo

2022, 26(4):  1-14.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.001

Asbtract ( 2760 )   HTML ( 101)   PDF (1076KB) ( 197 )  

References | Related Articles | Metrics

In the Internet age, more and more investors are beginning to express their investment opinions in online communities, especially in financial investment communities. The resulting massive financial text data has high research value. How to apply these financial text data has become the current research hotspots in the field of financial investment. This article explores how to convert investor posts in the Eastmoney Stock Forum into corresponding sentiment indicators, and form investor opinions based on this, and builds a portfolio model that considers financial text sentiment information under the framework of the Black-Litterman model. Specifically, we first use web crawlers to crawl the post text data of FTSE China’s A50 constituent stocks from the Eastmoney Stock Forum, and perform data preprocessing. Then, the sentiment indicators of the post text is extracted by using the dictionary method and the Naive Bayes method. Furthermore, three indicators of sentiment index, stock closing price and trading volume are taken as characteristic variables, and the random forest regression algorithm is used to predict the future return rate of stocks. Finally, the predicted future return rate is taken as the investor’s point of view, and is put into the framework of Black-Litterman model to construct a new portfolio model considering the emotional information of financial text. The backtest results show that the financial text sentiment mining portfolio model based on the Naive Bayes method has better performance.

Research on the regulation of online car-hailing based on stochastic differential game

YANG Mingge, SUN Lulu, LIANG Xiaozhen

2022, 26(4):  15-30.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.002

Asbtract ( 2698 )   HTML ( 14)   PDF (991KB) ( 196 )  

References | Related Articles | Metrics

The local government, online car-hailing platform and driver were regarded as a regulatory system to discuss the regulation of online car-hailing from the perspective of tripartite game. In decentralized decision-making without central government subsidy, decentralized decision-making with central government subsidy, local alliance decision-making and centralized decision-making, stochastic differential game models were built respectively to study the following indicators, such as the optimal degree of regulatory effort for the local government and the online car-hailing platform, the optimal degree of service effort for the online car-hailing driver, the expectation and variance of online car-hailing goodwill, the optimal benefit of the system members and system. Some important results were derived. i) Compared with decentralized decisionmaking without central government subsidy, all the above indicators increase in decentralized decision-making with central government subsidy. ii) Compared with decentralized decision-making with central government subsidy, in local alliance decision-making, the optimal degree of regulatory effort for the local government remains unchanged, while the other indicators all increase. iii) Compared with local alliance decision-making, all the above indicators increase in centralized decision-making.

A clustering-based surrogate-assisted evolutionary algorithm for expensive multi-objective optimization

BAI Fusheng, CHEN Jiaoling

2022, 26(4):  31-42.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.003

Asbtract ( 2840 )   HTML ( 11)   PDF (921KB) ( 181 )  

References | Related Articles | Metrics

A clustering-based surrogate-assisted evolutionary algorithm is proposed for computationally expensive multi-objective optimization problems. Under the framework of MOEA/D, the population is partitioned into several clusters, and the population subsets are formed via the neighbourhood of the weights. Then the radial basis function surrogate-assisted differential evolution algorithm is used to generate new solution points from the formed subsets, and the population is updated using the generated new solution. Numerical experiments have been undertaken on 7 DTLZ test problems, and the computational results indicate that the proposed evolutionary algorithm has advantages over the newly developed multi-objective neighborhood regression optimization (MONRO) algorithm.

Systematic tail beta and hedge of the market crash risk: based on a “safety-first” portfolio selection equilibrium model with the crash risk

LING Aifan, ZHU Jialei, TANG Le, JIANG Chonghui

2022, 26(4):  43-63.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.004

Asbtract ( 2752 )   HTML ( 14)   PDF (1076KB) ( 169 )  

References | Related Articles | Metrics

Recently, thousands of shares fell on the same trading date frequently happens in China’s A-share market. How to measure and predict the disaster risk when the market crashes are paid to the close attentions. To answer these problems, we establish the “safety-first” portfolio selection model with the market crash risk constraint, and get a crash capital asset pricing model (CCAPM) under the equilibrium condition. Combining the market beta, we construct a new systematic disaster risk measure, systematic tail beta, β^∆, and study its estimation approach. The empirical results using the daily returns of A-share market between 1995 and 2018 show that the β^∆ can effectively capture the tail comovement of the risk asset and the market during the market crash and boom. Especially, β^∆ has a significant positive impact on the tail returns of risk assets during the market crash. The H-L portfolios composed of the difference between High and Low β^∆ portfolios can obtain the significantly and positively average tail returns when the market crash occurs. These empirical results provide the important foundation to effectively hedge the market crash risk.

Block alternating proximal gradient algorithm for convex-nonconcave minimax problems

ZHANG Huiling, XU Yang, XU Zi

2022, 26(4):  64-74.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.005

Asbtract ( 2704 )   HTML ( 34)   PDF (817KB) ( 183 )  

References | Related Articles | Metrics

This paper proposes a single-loop block-alternating proximal gradient algorithm to solve block convex-nonconcave minimax optimization problems. In each iteration of the algorithm, the proximal gradient method is used to alternately update each variable in the objective function. We have theoretically proved that the algorithm achieves an ε-stationary point in O(ε^-4) iterations. To the best of our knowledge, this is the first time that a single loop algorithm has been proposed to solve a block convexnonconcave minimax optimization problem.

Scheduling with periodic due date to minimize the maximum tardiness

WAN Long, HUANG Xiaoli, MEI Jiajie

2022, 26(4):  75-86.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.006

Asbtract ( 2661 )   HTML ( 13)   PDF (815KB) ( 134 )  

References | Related Articles | Metrics

In the paper, we consider novel scheduling problems under single-machine, parallel-machine and two-machine open-shop environment. The due date is assigned to a job according to its completion time and the intervals between consecutive due dates are identical. Generally, the due dates are referred as periodic due dates (PDD). The objectives we consider are all to minimize the maximum tardiness. For single machine environment, we give an optimal schedule in polynomial time; for two parallel-machine environment, we prove this problem is NP-hard; for parallel-machine environment, we prove this problem is strongly NP-hard; for two-machine open-shop environment, we prove this problem is strongly NP-hard.

A class of spectral conjugate gradient method with sufficient descent property

LIU Pengjie, JIANG Xianzhen, SONG Dan

2022, 26(4):  87-97.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.007

Asbtract ( 2692 )   HTML ( 14)   PDF (1057KB) ( 152 )  

References | Related Articles | Metrics

The spectral conjugate gradient method is an important extension of the conjugate gradient method, and is one of the effective methods for solving large-scale unconstrained optimization. The designing for the spectral parameter is a critical work in spectral conjugate gradient method. In this paper, a new spectral parameter is given, and a new framework of spectral conjugate gradient method is established when the conjugate parameter satisfies a certain restrictive condition. Under the general assumptions and in case where the strong Wolfe inexact line search criterion to yield the step length, the new algorithm framework have sufficient descent property and global convergence. Finally, for the new algorithm framework, the existing conjugate parameter that satisfies the restrictive condition is selected, and the numerical experiments are done to compare the proposed algorithm with other potential algorithms, and the numerical results show that the established algorithm is promising.

Cartesian product graphs with crossing number two

WANG Jing, ZHANG Zuozheng

2022, 26(4):  98-106.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.008

Asbtract ( 2625 )   HTML ( 294)   PDF (867KB) ( 96 )  

References | Related Articles | Metrics

The crossing number of a graph G, denoted by cr(G), is the minimum number of edge crossings in all drawings of G. The research on the crossing number of a graph is an active problem in topology graph theory. Klešč and Petrillová characterized graphs G₁ and G₂ for which the crossing number of G₁□G₂ is two if G₁ is a cycle. This paper studies the necessary and sufficient conditions of G₁ and G₂ for which cr(G₁□G₂) = 2 if |V(G₁)| > 3.

The within groups and the between groups Position values on hypergraph games

HAN Jiayu, YU Zhiqiang, ZHAO Jiagui, SHAN Erfang

2022, 26(4):  107-118.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.009

Asbtract ( 2635 )   HTML ( 237)   PDF (884KB) ( 111 )  

References | Related Articles | Metrics

We study the additive decomposition of the Position value on hypergraph games. In 1988, Meessen considered the contributions of the links in the alliances and proposed an important allocation rule, called the Position value. By considering each conference in hypergraphs not only affects the benefit of the players in the alliances associated with it, but also affects the benefit of the players in the alliances that are not associated with it, we introduce the within groups Position value and the between groups Position value on hypergraph games to distinguish the components of each player’s benefit. We first give axiomatic characterizations for these two kinds of values. Secondly, we give an example to illustrate the within groups Position value and the between groups Position value. Finally, we propose an improved allocation rule by adjusting the proportion of the intermediary cost.

Counting the spanning trees for a class of self-similar planar graphs based on techniques from electrical networks

ZHAO Weiliang, QI Linming

2022, 26(4):  119-126.  doi:10.15960/j.cnki.issn.1007-6093.2022.04.010

Asbtract ( 2589 )   HTML ( 276)   PDF (839KB) ( 116 )  

References | Related Articles | Metrics

The Sierpiński gasket pyramid networks are the sketches of the Sierpiński tetrahedra which are the three-dimensional analogue of the Sierpiński triangles. Motivated by the construction of Sierpiński gasket pyramid networks, in this work we study the spanning trees of a new class of self-similar planar networks which has a very similar iterative generating method. By using the self-similarity and employing techniques from electrical networks, we obtain the exact analytical expression for the number of spanning trees of this kind of planar networks as well as the entropy of spanning trees.